August 9, 2026
Proof, panic, and chatbot tea
A partial digestion of the HRT counterexample
A famous math “impossible” just cracked — and the comments are obsessed with the AI angle
TLDR: A major math conjecture just got a counterexample, which is a big deal because experts had expected this version to hold. But in the comments, the real buzz is about AI’s role and the shared GPT chats, with readers treating the proof like equal parts breakthrough and backstage drama.
Math internet got a fresh dose of chaos after a long-standing puzzle, the HRT conjecture, took a surprise hit. In plain English: people thought a certain kind of nicely behaved function couldn’t produce a finite “copy-and-shift” relation unless it was something special. Now a new result says: actually, there is a counterexample — even in a very well-behaved case mathematicians had hoped was safe. That alone would have been huge. But the real fuel for the comment-section fire? AI was involved.
The mood in the community is a mix of awe, curiosity, and that classic internet side-eye. One camp is impressed that the authors seem to have handled the AI issue carefully, with the final proof written by humans and extra checks done the old-fashioned way. Another camp is clearly fascinated by the meta-story: not just “big theorem falls,” but “big theorem falls and now everyone wants to inspect the chatbot receipts.” That’s why one of the standout reactions was people immediately linking to a Reddit thread and to Tao’s shared GPT conversation, turning the proof into a community watch party.
There isn’t a full-blown flame war in the tiny comment sample, but the vibe is unmistakable: the theorem is important, the AI angle is irresistible, and the audience is eating up the behind-the-scenes drama. For a notoriously abstract math result, that’s about as close as you get to popcorn-worthy.
Key Points
- •The article reports a counterexample to the Heil-Ramanathan-Topiwala conjecture, including in the Schwartz-function case.
- •Prior positive results included Linnell’s theorem for shifts in translates of discrete subgroups of \(\mathbb{R}^2\) and Bownik-Speegle results for suitably super-exponentially decaying functions.
- •The new theorem states that there exist distinct points, nonzero coefficients, and a non-zero Schwartz function satisfying a nontrivial linear relation among time-frequency shifts.
- •The article says the proof was AI-assisted and also used a traditional numerical computation to verify one step.
- •At a high level, the proof reduces the problem to an eigenvalue problem and then to a vector cocycle problem via a vector-valued Zak transform on a denser lattice.